English summary for screening — check the original posting before applying.
Matsuo Lab at The University of Tokyo is recruiting LLM Development Engineers to accelerate research and development in understanding and controlling Large Language Models. This role involves contributing to major LLM research, including the development of models like Weblab-10B.
Must-haves
- Master's degree (or expected) in Computer Science or higher.
- Experience training neural networks on multi-node, multi-GPU systems (e.g., Deep Speed, Megatron LM).
- Experience in the full cycle of data processing, model training, and evaluation for generative language models.
- Ability to communicate and write in Japanese.
- Ability to read and understand English documents for information gathering.
Nice-to-haves
- Experience leading a development team.
- Experience releasing developed language models.
- Experience in reproducible implementation of technologies presented at international conferences.
Tech stack
LLMWeblab-10BMambaDeep SpeedMegatron LMRetrieval Augmented Language ModelsToolkenCameleon
Work style
Onsite in Hongo, Bunkyo-ku, Tokyo. Flextime system available.
Other notes
Salary range: 5.5 to 14 million JPY per year. Employment type: University of Tokyo Special Fixed-term Faculty Member.